2021 was better than the annus horribilis of 2020, but it was still a difficult year as Covid-19 continued to suppress travel and socializing, killed more Americans (and likely others globally) than the prior year, and mutated in strange ways (e.g. Delta and Omicron) to show the world that viruses don’t care about politics or our ineffective disease suppression or health care systems. It was another weak year for me as a reader, since I had a tough time focusing and was also working on some distracting side projects with generative GANS, crypto-art/NFTs, and moss art. Note that these books and papers could have been published in any year, but 2021 was the year I read them and gave my thumbs up.
BEST BOOK OF 2021: Doudna and Sternberg, “A Crack in Creation: Gene Editing and the Unthinkable Power to Control Evolution.” Gene editing will allow us to control plants, animals, and humans down to a cellular level. We have much of the technology today, but not the daring to aggressively use it, given how powerful and unknown it is. Doudna is a UC Berkeley prof and was one of the 3 (or 12, or hundreds of) people who helped create CRISPR. It will be a foundational technology for the next century, if we have the wisdom and care to use it properly. Imagine superhuman kids who live to 120+, with average IQs of 180, hardiness to most diseases, and other types of outlier abilities. CRISPR also connects the world of biology to computation and will allow us to use AI to design better cybernetic beings down the road. This book is fairly technical (at a minimum you need a college biology or genetics course), so if you want an easier read, I would suggest Walter Isaacson’s recent biography of Doudna. The closest comparisons I can think of are Schrodinger’s “What is Life?”, Smith’s “The Origins of Life”, Jablonka’s “Evolution in Four Dimensions”, and Stewarts’ “The Mathematics of Life.”
Edward Frenkel, “Love and Math: The Heart of Hidden Reality.” A touching book by a UC Berkeley mathematician who grew up in anti-Semitic conditions in Russia and had to sneak into math seminars that he was banned from. Also a wonderful, plain English introduction to the Langlands program, one of the most exciting research directions in modern mathematics.
Epstein, “Range: Why Generalists Triumph in a Specialized World.” A thesis that excessive specialization is bad for humans and that a generalist skillset is necessary in the modern world, with some great examples. It’s a complicated proposition, but my belief is that a T-shaped skillset, or a Π-shaped one (a broad/generalist skillset with one or two areas of deep specialization), is better than either/or. However, society needs both generalists and specialists, and often specialists are better rewarded (think neuro-surgeons and cardiologists instead of GPs or internists). I’ve seen this in many situations, from bond trading to elite ML programming teams. Often groups of hyper-specialists are complemented by a few generalists that bring everything together (though the generalists I’ve seen tend to Π-shaped and worked very hard to get there). Giving children a range of experience (lots of exploration) makes much sense (from intellectual to manual training, and exposure to different cognitive styles), along with nudging them to a Π-shaped life. Also, “specialization is for insects.“
Beeman, “Plain Honest Men: The Making of the American Constitution.” I enjoyed this book as a clear history of how the US constitution was made, but it made me angry. While the system the Founders created was better than any other government for at least a hundred years, it had numerous flaws and is completely inappropriate for the 21st century (despite two “re-foundings” after the Civil War and during the New Deal). Roughly 55 elite men, of which 25 were slaveholders, came from 12 states to spend a few months in 1787 debating a document and flawed system that still controls 350mm Americans today. The best of the document came from Madison, Wilson, Franklin, Morris, and Washington (as the chair of the convention and moral exemplar). Huge issues had almost no debate: expanding federal authority with the “Necessary and Proper” clause, who can wage wars and how to impeach officials, age limits, state rights, fundamental individual rights, etc. I both greatly respect the document and I reject the idolatry of the constitution that so many Americans and lawyers hold. We need dozens of new state conventions and a new federal constitution; otherwise, the creaky code of our basic law will be insufficient and the US will eventually fall apart. I recommend you pair this book with Rakove’s “The Annotated U.S. Constitution and Declaration of Independence” and Amar’s “America’s Constitution: A Biography.”
Haskel and Westlake, “Capitalism without Capital: The Rise of the Intangible Economy.” Wealth in the 21st century is increasingly intellectual and digital. The bit is mightier than the atom. An even in the world of atoms, we find that ideas, organizations, and processes matter more than the raw input materials. This book is a primer on some economic principles that are often ignored in Econ 101 but are central to the modern economy and polity.
Stone, “Amazon Unbound: Jeff Bezos and the Invention of a Global Empire.” Bezos created one of the greatest organizations ever, in human history, in 28 years. It’s likely that 2021 is peak Amazon, with only stagnation and decline ahead after Bezos’s resignation, but it’s hard to know. This isn’t just a great business scaling story – it’s about creating one of the most efficient and successful organizations in human history, albeit with many challenges and problems. I’d also highly recommend a book looking at the principles and mechanisms that make Amazon work, by two former VPs (“Working Backwards: Insights, Stories, and Secrets from Inside Amazon”).
Rosenwein, “A short history of the Middle Ages, 4th Edition.” It’s a pesky part of human history – what the heck happened before the Renaissance, Reformation, and Enlightenment in Europe, during the period from 300CE to 1500CE? The book centers around the delightfully “dark ages” of Europe, with some Middle Eastern history and a sliver of North African, though highlights for me were the empires after Muhammed, lords and vassals across Europe, the Mongol hordes, entrepreneurs/traders/merchants and the Hanseatic League, the Crusades, medieval churches and music, and the Black Death.
Moody, “Rebel Code: Linux And The Open Source Revolution.” As a history of Linux from 1993 to 2000, it is a detailed analysis of the most important open-source project ever. It opened my eyes to the history of open source and free software under Richard Stallman, but also Torvald’s own influence in shaping the movement and the project that powers most of the servers on our planet. It inspires by showing that engineers with an appetite for high-quality technical outcomes can create a governance structure that works, without gobs of money needed as incentives. It’s also an outlier, in that the most developed open source projects today are supported by large companies (React, Tensorflow, Flutter, Angular, PyTorch, etc).
Russo, “The Infinite Machine: How an Army of Crypto-hackers Is Building the Next Internet with Ethereum.” The best history of how a ragtag group of randos and hackers met in Miami and decamped to Switzerland to build the most interesting Web3 and crypto project – the only one to rival Bitcoin. The Ethereum Virtual Machine is a “world computer” on a blockchain, and is important. This is a messy history of a project that may become as important as Linux one day.
Russell, “Human Compatible: Artificial Intelligence and the Problem of Control.“ A book by an AI professor at UC Berkeley, who wrote one of the standard textbooks. He frames the major AI problem of the next century as the “control” or “alignment” problem – how to make increasingly larger AI systems serve the needs of humanity, democracy, and the broader wealth and welfare, instead of just privileging a small group or worse, turning against humans in ways we cannot predict. Supporting a billion-scale AI system has made me think a lot about Russell’s arguments. A nice book to pair this with is Nilsson’s history of AI, “The Quest for Artificial Intelligence: A History of Ideas and Achievements.”
Ferguson, “Empire: How Britain Made the Modern World.” A stirring defense of the British Empire and its moral balance sheet of debits and credits with respect to what the Empire did to both improve India and impoverish it. On balance, I found Niall a bit too positive on the empire, but it certainly did have many positive effects (like building railroads and schools). Part of me wonders: if the administrators of HK had 20-30 years to run India from 1947 to 1980, instead of Nehru and his corrupt, nepotistic family, would the entire sub-continent be in a better place (a stronger democracy with high HDI scores)?
Ambrose, “Eisenhower: Soldier and President.” The two great American Presidents of the 20th century were FDR and Eisenhower. Eisenhower had many accomplishments: managing the allies as Supreme Commander and winning WW2; nurturing 8 years of peace and material prosperity with balanced budgets and centrist policies; tamping down partisanship in DC; holding back the military-industrial complex from unthinking expansion and belligerence; keeping the saner FDR reforms like Social Security and the regulatory state; building infrastructure like the Interstate Highway System; increasing STEM funding after Sputnik; desegregating the armed forces and Washington DC. His great mistakes were in being too timid around school desegregation, punting on Cuba and Vietnam, and not cultivating a sane next generation of GOP leaders (the extremists and Goldwater came back in 1964).
Soni and Goodman, “A Mind at Play: How Claude Shannon Invented the Information Age.” Shannon was the brilliant mathematician who invented information theory, wrote about boolean logic and early computational biology, and did important cryptographic research for the US government. He was the Bell Labs superstar, in a place filled with stars, and is one of the scientists of the 20th century I think the most about (alongside Turing, Von Neumann, Feynman, and others). Shannon liked to juggle while on unicycles and used his spare time to build useless things, like an AI mouse to complete a maze, chess-playing robots, etc. He used mathematical tools to invest in the stock markets and did quite well.
Cline, “Ready Player One.” A fun and fast read – basically teen fiction about how the Metaverse could play out, with an entertaining plot, loads of 1980s American pop culture references, and a reminder that we need a decentralized metaverse and not a single company dominating it. This is much better than “Snow Crash” and on par with “Neuromancer” and “Rainbow’s End” (though not as good as “Hyperion”). Avoid the movie.
Stephenson, “Seveneves.” The premise here is that the moon is destroyed, and its fragments will destroy human civilization in 2 years. Humanity has that much time to send a small group of people, mostly hardy scientists and engineers, off the planet with enough resources to survive the destruction and (hopefully) eventually rebuild earth. It’s a dark book. Almost everyone dies in gruesome ways (like 99.99999% of people), and even the survivors are doomed. Honestly, I was in a dark place in 2021 when I read this book, but I ended up being deeply inspired. Even in the worst of times, builders and teams can come together to survive, endure, and make important things. I warn you that this book is full of hard science, clever robots, and a mixture of the most negative and positive takes on humanity.
Powers, “The Overstory.” An eco-novel where the central characters are people and trees. A melancholy book that got bitter at times. The prose was touching and delicate, and it will make you see trees and ecosystems in new ways.
Vendler, “Dickinson: Selected Poems and Commentaries.” The dean of Dickinson scholarship writes some personal essays explicating, and also adding complexity and ambiguity, to some of the greatest poems ever written by the genius of Amherst, Massachusetts. I once visited Dickinson’s home, a historical site, and it was slow and modest in a striking way.
Ishiguro, “Klara and the Sun.” A mother with a sick daughter buys an Artificial Friend, an android, to keep her daughter company. They become friends, and we start to ponder about what it means to be companions and be human. Subtle, slow, and touching.
TECHNICAL BOOKS & PAPERS
Clark, “Designing an Internet.” Clark was one of the dozen or so principal shaper-designers of the modern internet, on par with Kleinrock, Cerf, Kahn, Berners-Lee, and others. This book is mostly for people with CS backgrounds or a strong grasp of Computer Networking or Distributed Systems (I barely pass). In it, Clark presents hundreds of alternative designs for different ways the internet could have grown and evolved, with many of these ideas laying dormant as future protocols for Web3 and the Metaverse. So it’s a must-read for internet aficionados and builders of the future worlds, along with policy and regulatory nerds, who want to redesign the internet from the ground up.
Lapan, “Deep Reinforcement Learning Hands-On: Apply modern RL methods to practical problems of chatbots, robotics, discrete optimization, web automation, and more, 2nd Edition.” A short and practical book on building deep RL systems. Deep learning has already shaken the computing world: most AI surrounding you 24/7 uses deep learning, from your phones to laptops, Gmail to messaging and Zoom, smart speakers to thermometers, and much more. Reinforcement learning is a theoretical branch that, when combined with deep learning, achieved stunning results in the last 5 years with projects like AlphaZero, OpenAI Five, Alphafold, and AlphaDogfight (to name a few). As more of these systems drop into production, the world will get even wilder.
- Bommasani et al., “On the Opportunities and Risks of Foundation Models“: A paper by the entire Stanford ML department on the future of massive AI models, esp. multi-input and multi-task models, and their ramifications.
- Silver et al., “Reward is enough“: The world’s top RL experts suggest that these models are surprisingly simple in what they need.
- Radford et al, “Learning Transferable Visual Models From Natural Language Supervision (CLIP)“: A new ML model that learns how to see (computer vision) by reading text (NLP), connecting two of the biggest areas of AI research and setting a trend for multi-modal models.
- Weston et al. “BlenderBot 2.0“: An open source chatbot that builds long-term memory and searches the internet, while having some filters for safe chat
- Marsh, “Formalising Trust as a Computational Concept“: An old 1994 Phd thesis that became really relevant in the Web3/crypto age.
- Jumper et al, “Highly accurate protein structure prediction with AlphaFold [Nature]”. Predicting the three-dimensional structure that a protein will adopt based solely on its amino acid sequence.
Crypto white papers
- Benet, “IPFS – Content Addressed, Versioned, P2P File System“: IPFS, as a permanent, decentralized, storage system, combines many things I love: BitTorrent, a high throughput content-addressed block storage model, content-addressed hyper links, generalized Merkle DAGs, versioned file systems, blockchains, etc. All to make a distributed hashtable, an incentivized block exchange, and a self-certifying namespace. My favorite crypto whitepaper, and one of the better projects.
- Braidenback et al. “Chainlink 2.0: Next Steps in the Evolution of Decentralized Oracle Networks“: Chainlink brings APIs to blockchains, and connects off-chain (aka “rest of internet”) data with blockchains and crypto projects. Decentralized oracle networks are a great idea – the trick will be to get them to work reliably and scale.
- Ordana et al. “Decentraland“: This is a decentralized virtual world, a metaverse that no one owns. It’s a much simpler whitepaper than the above two, more based on ideas of ownership and governance versus lots of technical detail, but it was fun and idealistic.
I’d like to end with some wisdom from a man in a red sweater that didn’t make my list.
Rogers, “The World According to Mr. Rogers.”
“Competition. It’s a word that makes many of us very edgy, and it’s a situation that we have probably been living with since we were very small. For some people competition is a thrill, a stimulation, a challenge. For others, it’s a source of sadness and anger and apprehension. For still others, it’s a mixture of all those things. It’s not possible to go through life without competing. As one woman told me, ‘Competition is a part of our everyday life, whether we’re competing for a job, or on the soccer field, or for love.’”
“You rarely have time for everything you want in this life, so you need to make choices. And hopefully your choices can come from a deep sense of who you are.”